An in vitro method for investigating drug penetration across the cornea from acute doses has been developed. The donor (epithelial or tear) side of a corneal chamber similar to those used by Edelhauser and co-workers was modified with a reduced-volume insert so that the donor side volume approximated 6% of the receiver (endothelial or aqueous humor) chamber volume. Fluid was pumped through the anterior chamber to simulate the physiologic tear turnover in vivo. The receiver chamber was bubbled with oxygen:carbon dioxide (95%:5%) to aerate and circulate the fluid. We investigated the in vitro ocular bioavailability of several marketed ophthalmic drug formulations using this model (0.03% flurbiprofen, 0.5% levobunolol, 0.1 and 0.25% fluorometholone, and 1% prednisolone acetate). At 1 min after the dose administration, drug was eluted from the donor chamber at a dose turnover rate of 12%/min. In all experiments, a small percentage of the applied dose penetrated the cornea. Using this chamber device, the ocular bioavailability via corneal absorption ranged between 5 and 16% for solution formulations, whereas the ocular bioavailability for the suspension formulations was less than 0.3%. The major portion of the administered dose was recovered from the donor side effluent. This observation was consistent with findings showing low ocular bioavailability of ophthalmic preparations due to the rapid precorneal washout in vivo. The extent of drug penetrating as its corneal metabolite in the model correlated well to in vitro corneal metabolism rate constants. The new model opens several potentially useful areas of research into ocular absorption.(ABSTRACT TRUNCATED AT 250 WORDS)